Evolution of information during problem solving in inventive design

Sébastien Dubois (1) , Roland de Guio (1) , Gérard Vieux (2)

1 LGeco - Laboratoire de Génie de la Conception
2 TRIXELL [Moirans]

Inventive design implies solving problems for which no known typical solution does exist. Designing a new system means building a representation of a concept that could be recognised and validated as a solution. Which knowledge is useful to represent a specific problem? What is the nature of this useful knowledge? Which amount of knowledge is required to solve a problem? What is an adequate way to represent the problem in order to solve it quickly? Several methods exist with different approaches for problem stating. To be able to compare these methods, it is important to understand which kind of concepts are used during problem solving. In this article a descriptive approach of design is proposed. One method, ARIZ, extracted from TRIZ – a theory for solving problems of invention – is analysed and described through its knowledge management dimension. Thus, the way inventive problems are reformulated to be solved is analysed. To reach this point, one solved case study lead by the application of ARIZ methods will be considered. An ontology representing the used frames to reformulate the problems, which allows to categorize and to make measures on the knowledge during the solving process, will be proposed. Then the treatment of the information during the resolution will be analysed and commented.

Mots clés: Inventive design; ontology building; method evaluation; TRIZ

Domaines: Autre

Sébastien Dubois, Roland de Guio, Gérard Vieux. Evolution of information during problem solving in inventive design. International Conference on Engineering Design (ICED'07), Aug 2007, Paris, France. pp.NA. ⟨hal-00341006⟩